Hypex Ncore

Status
Not open for further replies.
So here in Amsterdam I should scale the supply voltage down to 50V or so :) (I am forever changing light bulbs, as the old supply network from the 50's is less than stable but nothing gets replaced until it blows up)

Offtopic:

My dad had this issue at his office som eyears ago. A UPS kept throwing some servers down at night, turned out the voltage dropped so much the UPS tripped and safely shut down the servers. The net-guys actually monitored voltage for a full three weeks before they decided to dig new cables in to the main distribution point.

Right. Back on topic ...
 
Disabled Account
Joined 2008
Ontopic:

I've noticed an other very nice side of the Ncore's sound (or lack thereof)...

There are lots of recordings that are not polished to sound pretty on everything (like most so-called "audiophile" recordings are). In other words, a lot of recordings may sound harsh, "shouting" or grainy. But wait... the Ncore's fabulous resolution manages to dissolve such recordings into its real content instead of just presenting it as one big lump of dirt.

I guarantee you that some will complain about this, mistaking midrange "shout" for dynamics. Which is utterly wrong of course.
 
Guys, sorry if I missed it browsing through the pages...

...but is there still any form of opamps on the ncore?

As I'm upgrading B&W 804S to the latest B&W 802 Diamonds, I might also get Ncore amps just to be 'sure'. My 'preamp' is a W4S DAC2.

On a sidenote, I did test my DIY amp (straightforward dual mono UCD400HG/SMPS400) on said 802 diamonds, right after listening to them on a devialet amp (hybrid AD) - they come pretty close to the devialet and had absolutely no problem driving those speakers LOUD. They were driven straight from an accuphase SACD player with variable XLR output.

Thx!
 
Ontopic:

I've noticed an other very nice side of the Ncore's sound (or lack thereof)...

There are lots of recordings that are not polished to sound pretty on everything (like most so-called "audiophile" recordings are). In other words, a lot of recordings may sound harsh, "shouting" or grainy. But wait... the Ncore's fabulous resolution manages to dissolve such recordings into its real content instead of just presenting it as one big lump of dirt.

I guarantee you that some will complain about this, mistaking midrange "shout" for dynamics. Which is utterly wrong of course.
So you don't experience some of the opposite; That bad recordings start to sound even harsher and more grainy due to more transparency?
 
So you don't experience some of the opposite; That bad recordings start to sound even harsher and more grainy due to more transparency?

That sort of thing commonly happens on replay systems which try to "sound transparent". It's a bit of a terminology thing. Some people use the word to mean "you get a barrage of details thrown at you", others use it to say "can't hear it do anything to the sound at all".

The first kind of "transparency" is actually a euphonic wow-effect which you can create by dipping the frequency response a dB or two in two places (600Hz and 2kHz) and optionally adding a small bit of HF distortion in the electronics (e.g. using all power FETs and no feedback). That produces what reviewers would describe as "startling clarity" and indeed such a system tends to overemphasize some of the things that bad recordings do badly.

But in the latter case (true transparency), when a recording sounds mucky, you only get the exact amount of muck that was in the recording, not artificially enlarged by the magnifying glass of an overzealous loudspeaker.

I knew all of this in theory until recently when the pieces in my active speaker design really started to fall into place. I now know it in practice too :) These days I can listen to literally anything on my main system. I no longer need a ghetto blaster on the side to listen to second rate stuff and make it listenable. I hear precisely what the recording does wrong but it doesn't keep unnecessarily drawing my attention to it. I can just get on with the music.
 
Last edited:
Well I've certainly never experienced the 'as the system gets better, the less I can listen to my poorly recorded stuff' phenomenon. If the recording is so poor as to sound bad, it sounds bad on any system reproducing it. If you reproduce it on a poor system you end up with poor system + poor recording. If you play it on a really good system it only sounds as bad as the recording is. Naturally you might hear just how bad the recording is, but it isn't going to hurt your ears any more then the poor system + poor recording is.
 
Bruno, I have been studying the new datasheet with a view to replacing the UCD modules with NCore modules in my multi-channel home theater amp and have a number of questions that I can't get from the datasheet or earlier postings in this thread.

1) Transformers. I use 500 VA toroid transformers per UCD400 module. However toroids let in more mains hash – is this a concern for the NCore modules? As the PSRR is 80dB (more than the 65dB for the UCD400) I assume it is not so critical to do any mains filtering? How good is the NCore PSRR above 1Khz as the 80dB is rated for <1Khz yet the UCD400 figure is for all frequencies.

2) Grounding. It is illegal in Australia to use a metal case without it connected to mains earth (or use a double isolated chassis). If using a balanced input with XLR and wired correctly for balanced, do you see any problems with earthing the chassis?

3) Distance between modules. I have UCD400 & 180 modules mounted vertically about 2 inches apart, and had no problems with EMI interference between modules. How more or less sensitive are the NCore modules to mounting close together, or mounting close to older UCD modules?

4) The round heatsinks are a problem for retrofitting Ncore modules to a chassis designed for UCD modules, as it is not always convenient to bold the NCore to the bottom of a chassis (as assumed by the datasheet). I have to mount the Ncore modules vertically, so I assume the best way is to make a L bracket that bolts to the circular heatsink and then to amp chassis sides (where the old UCD modules were mounted). Any other suggestions?

5) Heat. As idle losses are ½ as much as a UCD400, and efficiency is 93% vs 92%, an NCore amp should run cooler than an equivalent UCD amp. However not using an external gate drive supply the idle power supply usage is approximately the same as the UCD400 due to the on board regulator. My current amp case (with 2xUCD400, 3 x UCD180) only gets warm, never hot, so should be fine to retrofit NCore modules without additional heatsinking.

6) Apart from lower module heat dissipation, is there any other advantage to using an offboard gate drive supply?

7) DC protection. Is it possible to use the DC error detection on the NCore to replace an external DC protection circuit? There little information about the DC protection circuit (eg. how long it takes before it triggers, what is the minimum DC output voltage needed to trigger etc.)

8) How regulated / low noise does the op amp supply need to be for VSig? Any recommended power circuits to use?

9) I'm interested in your thoughts on matching IV stages to the NCore - Any recommended IV stages to use as input? I’ll be using a ESS9018 DAC but there are many options for IV. Given the effort to design a discrete buffer for the NCore, I’m thinking a discrete I/V would be a good match (although I’m aware that any reasonable IV would do).

10) Csup. I assume Csup is the minimum power supply caps to use, not the on board power supply caps which looking at the photo of the board are pretty small? I have 10,000uF per rail per module which should be fine.

Also please make more modules for the January production run, so they don't run out so quickly!

Thanks for your feedback.
Dean
 
Last edited:
@Dean
1. Our SMPS aren't regulated either, the output looks very much like that of a transformer based supply. So just keep the supply you already have, it'll do fine. The PSRR is rated <1kHz because it does degrade to UcD-like values higher up. This is because at high frequencies we're not looking at "true PSRR" but magnetic coupling between traces and that hasn't changed between UcD and Ncore.
2. No problem. So long as XLR pin 1 goes straight to chassis in all of your equipment, grounding has no effect on quality at all, that's the beauty of it.
3. The Ncore module is even less sensitive (about 20dB)
4. L bracket is fine.
5. Heatsinking that works for the UcD400 will be more than enough for the NC400
6. No
7. The nFATAL output triggers in a few tens of ms when the fault is full scale, and it will never trigger below 15V.
8. The regulators on the board are essentially the HxR12 circuit which has phenomenal noise rejection. Just rectify and smooth with a cap.
9. The choice for I/V should not be dictated by what follows, so I'm not going to dive into an I/V discussion. I'd only suggest that you stay away from the sort of minimalist discrete I/V's that are sometimes brandished about. If you design a discrete circuit, it must be because you can measurably beat IC based ones. Otherwise there's no point.
10. Yes we only specify the minimum. There's no maximum :)
 
... they come pretty close to the devialet and had absolutely no problem driving those speakers LOUD. They were driven straight from an accuphase SACD player with variable XLR output.

You're not only listening to the output stage then, as the Devialet has everything you can possibly think of integrated :). Cool piece of engineering, btw that Devialet. I haven;t been lucky enough to listen to it.

Any clear differences in sound between the devialet and the ucd?
 
Hehe sorry no, I explained badly.

I compared the 802 on Devialet, can't remember which SACD source, digitally with coax cable into the devialet.

When it was time to listen to the Hypex, we used the Accuphase as a source, because there was no balanced amplifier there, and it was the only way to listen with volume control.

The Devialet is exceptional, really. The room I listened in had a lot of issues, but I made the best of it.

The amps are very very close as far as I can tell, perhaps a smaller soundstage on the UCD, but that's all. Couldn't believe how commanding the stock UCD's are. In a way the UCD was more refined, but I lacked the "being there" feeling a little bit.

But honestly, it could just be "in my head" because I knew I was listening to a 12.000 euro D/A/Amplifier before. The difference is so small I couldn't possibly know/remember if I would have heard both solutions at different moments in time.

Anyway, back on topic - "to be sure", when it's possible I will build an ncore amp for these puppies. :)
 
That sort of thing commonly happens on replay systems which try to "sound transparent". It's a bit of a terminology thing. Some people use the word to mean "you get a barrage of details thrown at you", others use it to say "can't hear it do anything to the sound at all".

The first kind of "transparency" is actually a euphonic wow-effect which you can create by dipping the frequency response a dB or two in two places (600Hz and 2kHz) and optionally adding a small bit of HF distortion in the electronics (e.g. using all power FETs and no feedback). That produces what reviewers would describe as "startling clarity" and indeed such a system tends to overemphasize some of the things that bad recordings do badly.

But in the latter case (true transparency), when a recording sounds mucky, you only get the exact amount of muck that was in the recording, not artificially enlarged by the magnifying glass of an overzealous loudspeaker.

I knew all of this in theory until recently when the pieces in my active speaker design really started to fall into place. I now know it in practice too :) These days I can listen to literally anything on my main system. I no longer need a ghetto blaster on the side to listen to second rate stuff and make it listenable. I hear precisely what the recording does wrong but it doesn't keep unnecessarily drawing my attention to it. I can just get on with the music.
Good answer. That was my initial thought too.

It's seems especially appreciable to have a system with high transparency when playing recordings that are a bit "boxed in". It helps to hear whats going on and everything is served with more clarity.

StigErik:
"I'm afraid" your feedback about Ncore cost me about 25000 NOK. :smash:
 
Status
Not open for further replies.